Hapgood

For the past few months I have been talking about “choral explanations” and how they might transform our approach to OER. This is an outgrowth of my work on federated wiki and Wikity, but is a much more specific and immediately applicable idea. In fact, as I will show below, choral explanations are already in use elsewhere on the web: it is just that education has not yet made use of this pattern.

This post is my attempt to pull together in one place my thinking about choral explanations as a way of approaching educational materials: why it’s important, and what opportunities it provides. We’ll start by talking about how textbooks are currently produced, move on to new collaboration models we’re seeing on the web, walk through a specific example, and finish by talking about the long-term vision.

The Chum and the Ore

A little over a decade ago, award-winning author Tamim Ansary wrote an article on the way commercial textbooks get made.  He had been tasked to help an army of writers to write a book “from scratch”. The process turned out to be less than inspiring:

Sounds like a mandate for innovation, right? It wasn’t. We got all the language arts textbooks in use and went through them carefully, jotting down every topic, subtopic, skill, and subskill we could find at each grade level. We compiled these into a master list, eliminated the redundancies, and came up with the core content of our new textbook. Or, as I like to call it, the “chum.”

After several more steps (including adding in whatever hot new pedagogical theory was in vogue) the writers got down to writing. Sort of.

Finally, they divide the outline into theoretically manageable parts and assign these to writers to flesh into sentences.

What comes back isn’t even close to being the book. The first project I worked on was at this stage when I arrived. My assignment was to reduce a stack of pages 17 inches high, supplied by 40 writers, to a 3-inch stack that would sound as if it had all come from one source. The original text was just ore. A few of the original words survived, I suppose, but no whole sentences.

Though academic publishing in higher education is a bit different than what is described in some of his article, contacts I have in the industry tell me it’s not different by much.

From Voice into Pulp

Key to the process that Ansary describes is an obsession with eliminating individual voice, originality, and viewpoint.

To avoid the unwelcome appearance of originality at this stage, editors send their writers voluminous guidelines. I am one of these writers, and this summer I wrote a ten-page story for a reading program. The guideline for the assignment, delivered to me in a three-ring binder, was 300 pages long.

As Ansary describes it, it’s an expensive and difficult process to coordinate. Getting a coherent, single voice out of multiple contributors is difficult. Getting it “objective” is even harder.

Oddly, when I read the Ansary’s piece I was reminded most of Wikipedia. Oddly, because Wikipedia is often held up by the publishing community as a cautionary tale when it comes to open materials. But this process – trying to get dozens or hundreds of voices and past snippets of writing pounded into a coherent whole – is what Wikipedia does as well. It’s both the key to its success and the cause of its current malaise. Ten of thousands of hours are spent editing Wikipedia’s top articles, but for the most part they aren’t spent coming up with new ways to explain things, or updating articles with new research. They are spent are the never-ending pulping out of voice, perspective, bias, and differing opinion about what belongs in the article. Only a sliver of writing – even good writing – makes it through.

And the writing that does make it through? It’s often uninspiring, dry, and voiceless.

As I’ll make clear later in this article, I believe Wikipedia does some things very well, and remains one of the great achievements of networked culture, just as the modern, bureaucratic publication process followed by the textbook companies was a great 20th century achievement. (I’m not kidding here: it was).

But two questions arise when we look at these processes. First, are there more efficient ways to put together articles than this endless process of pulping things into a single narrative? And second, is it possible that those ways might make for a better experience for the reader as well?

I believe the answer to both questions is yes. I believe that by moving past our romance with the textbook as a single authoritative voice that we can ultimately produce more effective works that are also more amenable to collaborative production by both faculty and students. And I think (or perhaps hope) that producers of Open Educational Resources can lead the way on this, and renew their commitment to open pedagogy in the process.

The Problems of Textbook Publishing Are the Problems of Wiki

I spent many years, from the early aughts onward,  promoting wiki as the prime example of networked collaboration, to varied success. Then several years ago I came across a video of Ward Cunningham, the inventor of wiki, saying he may have got key elements of wiki wrong.

The problem, said Ward, was that wiki was a relentless consensus engine. And for certain things (e.g. encyclopedias) that might not be a bad thing, but as a way of working it had its drawbacks.

First, let’s acknowledge the benefits. Works that aim for what we might call the “encyclopedic voice” can be more accessible to newcomers. They rely less on reader background knowledge, and can be much more helpful to a novice than being dropped into a cacophony of different, competing voices. Works produced by either Wikipedia-like consensus or the editorial pulping that Ansary describes can be particularly useful to students for some aspects of their course, providing the student a map of the territory before pushing them to delve into to a particular journey.

The drawbacks, however, should be familiar to anyone who has worked in a consensus publishing environment, whether on a textbook, a Wikipedia article, or a policy document. Consensus can be off-putting to contributors. It often suppresses important minority viewpoints. The process of document-by-consensus tends to value the bureaucratic over the human, and push people into spending  more time on edit wars and turf-defending than on production of knowledge or new insights. And in the end the work that is produced is often acceptable to everyone but exciting to none.

Above all, encyclopedic voice is expensive: expensive in terms of time involved, coordination cost, and toll on the production community. No matter whether one finds themselves embedded in Wikipedia’s headless bureaucracy or a publisher’s 300-page process, when multiple voices must be merged, the cost of production jumps sharply. Well-known and much-feared behaviors emerge: bike-shedding, edit wars, and the like. Once consensus is achieved, even small edits threaten to undo the hard-won agreements, making fluid evolution difficult. Over time, the result is often an unexpected rigidity that actively suppresses rather than spreads community knowledge.

Ward looked at these problems in 2011 and proposed a new direction for wiki, one he termed the “chorus of voices.” The idea of the chorus was that a wiki page’s title could form a hub for a number of individual, personal takes on a single idea, the way a hashtag can form the hub of a conversation on the web.  In the chorus, wiki editors don’t edit a single page: each editor creates their own version of the page, often out of the materials of previous pages (through a process called “forking”). Clicking on a link to a page on “The Causes of the Cambrian Explosion”, for example, would deliver to you one version of that wiki page, but make you aware of multiple other attempts to explain the same thing, often based on other versions of the same thing.

Part of the idea was to do here what processes like git had done for software: focus people on creating new work and fixing existing work rather than arguing about small points or winning edit wars by exhausting the opposition. But the more radical piece of the vision was making peace with “the chorus”; understanding that the “meaning” of a given page in wiki was the intersection of all the work individual authors had written against that title.

This would allow a lot of the personal perspective that gets pulped out in a traditional wiki process to stay in. The marine biologist’s take on the Cambrian Explosion could focus on what they know best without having to be reconciled with the eye experts’ take on it, in which the evolutionary period is seen primarily as being about the emergence of vision. And sure, the anti-evolution set could write their own page, denying it ever happened, but in writing it as a separate work they would not exhaust the editors trying to make a scientifically accurate page with endless flame wars. In turn, this freedom from having to incessantly defend obvious revisions would attract new editors.

Over time, thought Ward, people would adopt or adapt the most useful pages, in effect creating an evolutionary process of their own, where the strongest pages survive not based on who has admin rights or the most time to babysit pages, but on which versions of pages were most useful to a given community.

I ended up working directly with Ward for over a year and a half on educational applications of his new approach to wiki, commonly known as “federated wiki”.  The experience ranks as one of the most influential of my life.

Together, we tried a series of experiments with writing this form, both in communities and in personal work, and in the process discovered a number of interesting patterns. And working in this way – seeing collaboration in this choral, almost upside-down way – made me attentive to things I might have otherwise missed. Subsequent work on a Shuttleworth-sponsored project called Wikity honed that radar further.

And so it was when I watched Joel Spolsky’s presentation to the Wikimedia foundation last fall — I saw more in it than I might have three years ago. Spolsky, in case you don’t know, is co-founder of Stack Exchange, a wildly popular “question and answer” site.  And in his Wikipedia talk he mentions that when “active editors” were defined in comparable ways, Stack Exchange (founded 2008) had actually surpassed the number of active editors found on English Wikipedia (founded 2001). And unlike English Wikipedia, Stack Exhange was still growing.

Wait, what?

As I thought about the similarities between Stack Exchange (formed 2008) and federated wiki (formed 2011), It occurred to me that Stack Exchange and Federated Wiki were in fact part of a broader movement in how collaboration was now happening in communities, and that it was time to bring this more general process into the OER movement.

Choral Explanations

To get at why the StackExchange model is so interesting requires a bit of a detour into how question and answer sites used to work.

Older sites (e.g. Yahoo answers) were essentially transactional. A person with a question would pose a question, and the answers below the question would respond to it. Eventually, the original poster of the question would select an answer as sufficient or best.  The question would be closed and people move on.

These older Q&A sites were simple variants of general forum architecture. And they get good results occasionally, but it also have the sort of problems a forum runs into — they tend to produce answers that look more like replies than generalized explanations.

For example, here’s some of the bottom answers to the question “How can a person increase their chances in a lottery?”

Again, what’s key here is the transactional nature of it. These answers do not read like wiki entries. They read like responses, and the reader must read them as they would any forum. Additionally, there is no real pooling of effort here – this question is asked on Yahoo dozens of times but those answers don’t form a chorus as much as a cacophony of scattered events:

Sites like Quora and Stack Exchange turned this process on its head. Instead of envisioning the Q&A site as a single-purpose forum, the new breed of Q&A site sees the model as half-wiki/half-forum. The question asked is analogous to the title of a wiki page; it’s not transactional but communal. A question like “How can a person increase their chances in the lottery?” is the place where the community will store their collective knowledge on that point, and it is not owned by the person who asked the question, but by the community itself.

On Quora, in fact, the question can (and often is) edited by the community for clarity, and on Stack Exchange posters who pose badly formulated questions are pushed by moderators to reformulate their question in ways more beneficial to the site. Duplicate questions are shut down, just as duplicate wiki pages would be, so that effort can be pooled. The original poster of the question has no more power than any other user to rate specific answers more useful than others or to close the question. And as with wiki, answers posted are meant to be complete answers, not lazy responses to or discussions with the original poster. Each answer is also self-sufficient (a pattern I have termed elsewhere as “hospitable”).

Posting a question on these sites is really not about starting a conversation at all. It’s saying “Let’s gather our community knowledge on this particular issue,” just as one might do with wiki.

Unlike wiki, however, individual control of writing is preserved, and multiple unique passes at a subject are appreciated. And big questions get a lot of passes. Here’s a snapshot of a few of the sixty-eight responses to Quora’s question of why many physicists believe in a multiverse.

The most fascinating thing? Unlike earlier sites, it’s not about the best or first adequate answer. People looking to learn a theory or a skill find seeing the multiple explanations a benefit. Since each response takes a different approach to providing an answer, the reader can read multiple explanations that get at a subject in different ways, at different levels of complexity. Some are nuanced, and some are ridiculously simplified. Some exercise metaphorical thinking, others dive into math, others illustrate with diagrams.

And this approach – multiple routes into the same concept for the learner – is supported by the research. There are no “learning styles”, as we know – no “kinesthetic learners”. But lost in the discussion about learning styles is the research base that shows that most students benefit from multiple approaches into a subject using a variety of styles. Segregating students into different style groups has no effect, but teaching in a variety of ways is quite effective.

The multiple explanations help in other ways too. Like most other users of Quora or StackExchange, I find that the tenuous understanding gathered reading an initial explanation is slowly solidified and clarified as I read subsequent explanations. In fact, this process –  reading multiple treatments of the same issue to add nuance to understanding – is a best practice for learners, accommodated and encouraged by this format.

There’s some important caveats here. The writers do follow community norms in their answers, and importantly, these explanations are presented as parallel attempts to answer the same tightly defined question. This allows a reader to quickly scan and process them in a way that is not possible when skipping through the results of a web search, for example.

Additionally, both StackExchange and Quora pay a lot of attention to formulating questions at the right level of specificity for their particular audiences. On the original Stack Overflow, Joel defined the ideal question as “I got this far with the code I wrote, but I can’t work out the next step.” Questions that ask things too far from a specific problem (“Is Python a better language to program in than Perl?”) get quickly deleted by moderators. Quora norms are different but just as particular about the types of questions that can be productively asked.

But largely what both these models represent is something like the chorus of voices Ward has been advocating – people who control their own answer, but place it in relation to other explanations on the site. For Ward, what ties it together is a page title. For Stack Exchange what ties it together is a question. But these seem to be part and parcel of a larger trend, one that split the difference between individual voice and community needs in a new and elegant way.

And the result? The site has quickly become one of the most popular and useful sites on the web.

Imagining Choral Explanations for OER

Why is all this important? Because for years we have imagined (or at least most people I know have imagined) an approach to OER production that looks like wiki or open source code production. In this model many people work on and gradually improve a small set of crowd-sourced textbooks. Maybe they create variations for their own purposes, and maybe some of their changes make it back to the mainline content. Maybe you end up with a very different version of the textbook you are using, but you ultimately end up handing the student one version of that textbook. We’ve seen models like this in WikiEducator, Connexions, etc.

But even though these methods are open and stigmergic, in these models we risk replicating the pulping process that Ansary describes. Even as we revise, remix, and redistribute, we still come to a work that is mono-tonal, bent towards a single voice, and often as soulless a voice as the ones Ansary laments.

I understand why that is, and why we spend time on this effort. If you’ve ever read a textbook where the voice was not hammered into something relatively uniform, you’ll know that it can be a painful and frustrating experience. The hard work that legions of former English majors do to turn these works (both commercial and open) into a unified whole should not be underestimated or go underappreciated.

But is it possible, just possible, that by adopting a different textbook model we could reduce the effort needed to produce such works, while making a product that is more effective for our students? And maybe even a product that allows them to easily contribute to its production?

First Step: Textbook Core as Operating System, not Application

To do this we’d have to start be reimagining what the core textbook looks like.

It’s interesting that Ansary, in the latter part of his article, doesn’t want to throw away the encyclopedic style of the textbook completely. The “view from nowhere”, as Thomas Nagel has called it, has its applications. People don’t buy a textbook on economics expecting to get solely Paul Krugman’s thoughts on economics or buy a physics textbook to get “Things Some Random Prof Thinks Are Important in Physics” – they buy a textbook specifically because they are interested in getting a more global view of a discipline than any one person can provide.

The problem is that we try to cover all the material in a course in this manner, when in reality most subjects only need a skeleton of ideas.

In his conclusion, Ansary gets the metaphor right:

In content areas like history and science, the core texts would be like mini-encyclopedias, fact-checked by experts in the field and then reviewed by master teachers for scope and sequence.

Dull? No, because these cores would not be the actual instructional material students would use. They would be analogous to operating systems in the world of software. If there are only a few of these and they’re pretty similar, it’s OK.

People have, of course, proposed this “textbook as OS” idea before, and I can hear some of the groans of the old-timers even as I propose this (for the record, I’m an old-timer too). It’s pretty hard to do “textbook as OS” when most faculty review textbooks by checking the table of contents to see if their three niche subjects are covered.

But in this case, I’m not proposing that material be removed from the textbook, but rather that It be separated in to two separate tracks, an encyclopedic, carefully sequenced core and a marketplace of choral explanations.

Here’s a glimpse of how that might work, and why it might be more effective for both students and textbook producers.

How Does Water Get Up That Tree and What Does It Carry With It?

For people who say that we don’t need educational materials or textbooks, I like to remind them of what it is like to be a student. (I have a sneaking suspicion that many people who complain that formal educational materials are not necessary have not tried to learn things outside their domain in a long while).

Here’s a paragraph expressing some of what you’re expected to understand to pass your biology class, taken from a model answer in a workbook. If you’ve stayed with this post this far, try and make an honest attempt to read the whole thing; we’ll be using this subject as our example for the rest of the post:

(Short tangent: When I write articles like this, readers often remark that I should use simpler examples that are more accessible to the general public. They point out it’s not a great idea to make your reader feel dumb by pushing them to slog through something like the above.  But what I’ve found is this – if we are not honest about the difficulties of comprehending dense technical prose we continue to end up with facile solutions and rhetoric around the issue of educational materials. So if reading that hurt your head or bored you out of your mind, well, I’m afraid that was the point. Welcome to being a student!)

As you’ve noticed, the sample paragraph is a bit dense. It was pulled from a set of model answers to workbook questions, and represents what students should know about water potential by the end of the chapter. But it requires that students master quite a bit of material even to effectively read it, never mind produce it.

To deal with this issue, what we generally do in textbooks is attempt to expand that compressed treatment, introducing the conceptual dependencies in sequence, referencing what the student should have learned before, and tying it to the new knowledge. So, for instance, we’ll take the brief mention in this paragraph of solute potential, and recognize that we’ll need to explain that in the text as part of our water potential explanation. And what we try to do there is come up with the “best possible” explanation of solute potential, often hammered out after reading how ten or twenty other textbooks and three contributing authors have explained the idea.

Here is the OpenStax Introductory Biology explanation of solute potential. Again, for maximum understanding of why choral explanations are necessary, please read the text below:

These are hard concepts, and the OpenStax copy here does its best to explain the concept in a concise but approachable way. It even adds a helpful diagram to demonstrate the relationship between solute and water flow:

But that’s it. That’s all you get. It may be the best explanation, or the most concise, but it’s also your only one. So what do you do as a student if it doesn’t work for you?

Well, you probably read that explanation again and again, and hope you understand it. If you’re an adept autodidact, maybe you’ll seek out some help on the internet, sorting through a maze of questionable and often confusing Google results. But if you’re not that autodidact, you’re stuck with this “one best explanation” the textbook has decided to provide you.

But why? How does that make any sense at all?

The idea of choral explanations in OER is that the textbook becomes an operating system on which multiple parallel community-provided explanations run. From the student perspective, the text branches off into multiple available explanations of the same concept, explanations authored individually by a wide range of instructors, researchers, and students. You can keep reading until you find the explanation that makes sense, or you can start with simpler explanations and work your way to nuance. (In the humanities and social sciences there are other more complex configurations that could be used, but we’ll leave those aside for the moment)

Here’s a mockup of what textbook as operating system might look like in practice:

When clicking through those links, you would come to Quora or Stack Exchange-type page where multiple people would take their crack at explaining the concept.  Unlike wiki-style production, each person explaining would get recognition for their work in the form of an avatar, and perhaps even a Quora-like bio which explains the source of the person’s expertise in a blurb (which could be anything from “Nobel Prize winner” to “A student who likes to explain things.”).

Here’s a quick mockup I made last month on mitosis which borrows directly from Quora’s design. What I want to give here the sense of the experience: bylined items with different approaches, and just as with Stack Exchange or Quora, you keep scrolling down until you start to understand.

You can imagine the same approach with a question such as “What is solute potential?” Or “Why is solute potential important?” (We’ll show an example of this in a minute).

Stigmergic Production

What about the production side of the equation? How do choral explanations fit into a new model of producing texts? Part of the answer can be found in the concept of stigmergy.

David Wiley has argued that stigmergy is the future of OER production (2004, 2016). He makes the case that coordinated, central work in OER publishing is too expensive to form a long term solution to most of our educational material needs. In the stigmergic pattern, what we hope to see is people creating the materials we need with no centralized governance. Collaboration comes about as a result of people pursuing individual aims in an environment that provides feedback to individuals on where profitable opportunities for work are.

What does this mean in practice? As an example, take Twitter. No one at Twitter says “these will be the hashtags for today”. Rather, someone tweets a message using a hashtag; other retweet it, or adopt the hashtag which exposes the hashtag to others, and so on. Eventually the hashtag trends. In stigmergic systems the environment is altered so that the work that one person does leaves “traces” — in this example, retweets, trending hashtags, visible omissions — that can direct the work of future workers and float the best or most interesting work to the top.

The attraction of stigmergic production is that by making participation lightweight and providing hints about work that needs to be done rather than control structures, people can produce coherent works without explicit coordination, usually at a lower cost and with greater variety than traditional systems.

Wiki is often cited as a stigmergic medium: since anyone can edit and errors and omissions are publicly visible, to use a resource like Wikipedia is also to discover the work that needs to be done. One person looks for a page or clicks a “red link” and doesn’t find it, and writes a stub article. A second person finds the stub and expands it, with errors. A third person finds it useful, but sees it needs a grammar cleanup and an error check and fixes it. A fourth person notices some “red links” in it that point to non-existent pages, goes and creates new stubs for them, and the process starts again.

For a long while I thought that this more traditional wiki use would be the route that stigmergic OER production would take.  In wiki, work tends to form around a set of articles that are slowly improved as they are used.

There may still be a central place for wiki-like production in OER, but newer choral styles of explanation provide an alternate means of production that has less overhead around consensus and synthesis. At the same time the process is still stigmergic.

Stigmergic Production Using Choral Explanations

Our  process begins with the writers of the core textbooks who identify initial productive questions and link to them in the text.  As shown above, these questions are chosen based on their relevance to the lesson at hand, and known student misconceptions:

From the links faculty can see what questions had many explanations and which were relatively uncovered. From there, if they want to contribute an additional explanation it would be as easy as clicking one of the questions while logged in.

When looking at a set of explanations in answer to a particular question, a faculty member could choose which explanations to show to students in a default view – over time the most used explanations could float to the top. Other criteria could be ranked as well, based on student and faculty feedback:

Faculty who do not wish to choose which explanations show up in the default view could simply choose to show questions with a given rating or above, or a set of explanations approved by an individual they trust.

If faculty find that the explanations are inadequate, or believe they have a unique way of explaining that is not covered, they can write their own explanation, which can then be used by other faculty. And if the textbook provided prompts aren’t sufficient, they can always create a new prompt.

What sorts of explanations might they produce? Maybe they film a video of themselves showing osmosis through a membrane in a beaker. Maybe they write up an example of how leeches and slugs react to salt to bring home a point about osmotic pressure that helps explain solute potential. Maybe that’s the same story they’ve used time and time again in their lectures, and they just know for some students it just works:

You’ll note a couple of things about this explanation. The first is that it’s not going to work for every student. For students who have seen a slug shrivel, this may be the perfect hook to hang this concept on. For ones who haven’t it’s meaningless.

It’s also at a relatively simple level. It doesn’t go into the chemistry of high and low potential, and explain how hydrogen bonds are behind the difference in potential – when salt binds with hydrogen atoms in distilled water, it consumes some of the energy of that water, taking it from a high-energy state to a lower one. But we don’t get into the math in that one.

But over time a library of explanations are created. Some use slugs as examples, some use dried flowers. Some build 3-D simulations, some show video of experiments. Some have mnemonics, some have diagrams. Some go deep into math and some stay out of the math.

As teachers decide to include or not include various explanations for their students, the highest teacher rated materials float to the top. As students read the explanations and rate them as helpful or not helpful the system builds a profile of explanation authors the student seems to prefer and suggest the student “subscribe” to that author, put their material to the top of the student’s view.

Over time, teachers find the best explanations and students find the explanations most targeted to their background, needs, and sensibilities. And if what floats to the top doesn’t work students can keep scrolling down until they find something that does.

If you’ve been following along and reading the pieces on solute potential, you’ll notice one final thing. By coming at the question a different way, the core explanation probably seems a bit more accessible at this point. Does this sentence start to make some sense?

It does, right? Now that you have the example of the slug, what seemed gibberish before starts to become more readable. My guess is that if you read a couple more explanations taking different approaches it would make even more sense. As each explanation takes a different angle on the explanation, novices are able to bootstrap themselves up into an understanding of the denser prose.

(This relates to the way in which choral explanations supports true personalization in a way most modern systems don’t which I’ve written about here).

A Chorus of Student Voices

As we look at the production process it becomes clear that this is also a way to accomplish a long-standing goal of the OER community: to bring students more fully into the production of OER. (The attempt to bring students into OER production at scale has a long history. For some of the history and current directions of this effort, see David Wiley’s recent post on “renewable assignments“).

If having students produce educational materials sounds like a distraction from the work the students should be doing in class, it shouldn’t. Consider the learning objectives for the chapter we’ve been looking at, for example:

When we look at these objectives, it becomes clear that the most fitting artifact to demonstrate competency is in fact an explanation. This is why the dream of a larger system of student produced OER has been so  attractive over the years – oftentimes the tasks that are the most authentic have a poor match with learning objectives, and the tasks that match with learning objectives often are inauthentic. Student-produced OER has the unique advantage that real, impactful work will have a close match to course objectives, almost by definition.

Using the Internet to publish student produced OER is not new: students have been involved with production of web-published educational materials many times before. The usual mode has been wiki or similar technology.

Most of these web projects were accomplished through running special class wikis that aim to cover a subject thoroughly in a separate site. While this allowed students a lot of freedom in the writing process it came with drawbacks. Class sites, when not promoted, often had little to no audience. Also, from the moment they were completed they began to decay. The story we told students – that they were contributing to the sum of human knowledge by publishing on the internet – was often belied by the state of these sites a year or two after the class – stagnant, often hacked and unseen by others. Additionally, finished sites often occupied an awkward middle space – without previous work to extend, these sites spun up from nothing often couldn’t cover enough material to be a resource of note, at least in the space of a semester (one thing I and many others have tried in order to mitigate this is running sites over several semesters, with each subsequent set of students extending the work of previous students, which has its own set of difficulties).

More recently a movement has grown up around having students contribute to Wikipedia pages. This solves many of the problems mentioned above: the use of Wikipedia guarantees student work will be seen: Wikipedia is, after all, still one of the top 20 destination sites on the internet, and still the source for most introductory subject information. It also puts the maintenance in the hand of someone else, and allows students to integrate their knowledge into a larger work.

Yet two major problems occur repeatedly with this model. The first is the problem of other people. Editing on Wikipedia, students spend a disproportionate amount time defending their changes to other Wikipedia editors. Sometimes this is pedagogically helpful – a student being pressed to find a source to support a claim made in the text. More often, though, the time goes to the pulping and blanding process, the removal of any individual style, insight, or personal experience the student might bring to the piece.

The other problem is even more difficult to overcome. If we look at those outcomes again we’ll see objectives like “Describe how water potential, evapotranspiration, and stomatal regulation influence how water is transported in plants.” But those explanations are made in Wikipedia on a few pages at most, which means that only a few students (if any) will have the chance to explain these things on Wikipedia from scratch. Other students will wrestle not with explaining core concepts in a unique way, but in the more mundane work of polishing old articles, a process far more about the use of language than about biology or physics.

And it’s not just a Wikipedia problem — any scheme which encourages students edit and revise a few central articles is going to have the same problem.

Choral explanations provides for a third way that promises some of the benefits of wiki without the drawbacks. As students use externally provide explanations in their class, they’ll be asked to rate them for helpfulness, depth, and completeness. But they’ll also be looking for concepts they believe that they can explain better or more fully.

Student teams can then get together and put together an explanation. Maybe one team figures out how to start an explanation on solute potential by showing a couple clips of people stranded at sea drinking saltwater in the movies, and moves on to show why you can’t grow a fresh-water plant in salt water.  Maybe another team wants to do a piece on how solute potential explains why salinization of soil is such a big problem in arid climates, and link it under the “Where can you see solute potential at work in daily life?”

Students create these pieces alongside the other explanations in the system in a private space. Because the model is a “chorus” and not a single central work, there is always room for students to bring something new and meaningful to the process (unlike Wikipedia, where mature articles offer little opportunity for additional contribution).

At the same time the work can be meaningful. With student permission, faculty can submit student work they deem sufficiently accurate into the central pool of explanations. These explanations, in turn, move up or down in prominence based on their usefulness to students. But in the context of student produced OER, they also serve as models to students of what sorts of things they could do on their projects, truly creating a system of renewable assignments.

Conclusion

As Open Educational Resources move into the mainstream, producers of these works must decide whether they wish to replicate the standard textbook process and format or explore new options in presentation and production. The choices we make now determine whether we will merely reproduce traditional textbook publishing at less cost or use this opportunity to leverage greater pedagogical change.

Yet even when publishers are committed to change, the nature of the model matters. Over the past twenty years we’ve learned a lot about collaboration in traditional wiki environments. Traditional wiki approaches address some of the long term needs of the open education community, but can also replicate unhelpful aspects traditional textbook publishing.

Choral Explanations is presented here as an alternative mode of working together on educational resources, one that allows collaboration and promotes coherence while preserving unique perspectives and facilitating easier community contribution. Additionally, in the way the model mimics differentiated instruction, it may provide a better and more personalized experience for the student as well.

This post is not meant to be a final blueprint of how such a scheme would work, but rather is an attempt to open up a conversation about how we might proceed in building an environment that balances individual control and voice with the pooling of effort communities require. A lot has happened since we began this fight to open up educational resources. As we suddenly find ourselves on the winning side of history, it’s important that we address these issues in serious, critical ways that take stock of what has and hasn’t worked in the past, and design these systems based on what we know in 2016. We have a massive opportunity in front of us: let’s not squander it.